Video: 10 Minute IT Jams - An update from Pixis with Hari Valiyath
Pixus wants to make artificial intelligence easier for marketers. The startup's co-founder and chief business officer, Harley Val, says the company's codeless AI infrastructure is breaking down traditional barriers for brands hoping to scale their marketing and sharpen their decision-making.
Pixus is positioning itself as a "co-pilot for performance marketing teams". Val explained, "So when you run digital marketing campaigns on Meta, Google, TikTok, all these ad networks, there are levers that you control manually, which is bid, budget, audiences and creative. We basically have an AI that automates all these four important functions of running a digital marketing campaign efficiently."
The company's main tools focus on automating audience targeting, creative content development, and optimising bidding and budgets across different advertising platforms. "The best part, all three systems are interconnected at the back end, and it keeps on learning from the data it gets from all the clients," he said.
A standout feature for Pixus, Val said, is its cross-channel optimisation. "Shifting dollars between Meta, Google, TikTok – all these ad networks are black boxes as we know. So let's say if an e-commerce company wants to optimise for ROAS, today the decision making of allocating dollars between Meta and Google or TikTok is done manually. We have an AI that plugs into their attribution platform data and automates it. That, I would call a flagship feature," he explained.
Asked about the problem Pixus aims to solve in the marketing AI landscape, Val pointed to complexity and a lack of transparency. "There has been a lot of myth that AI is a black box. There is no transparency – it is difficult to integrate. The integration time can be anywhere from two to three weeks," he said.
Pixus's answer is a plug-and-play Chrome extension that overlays onto common ad management tools, requiring no code or technical integration. "There is no heavy lift required at all in terms of integration – it is the same Ads Manager [marketers] log into on a day-to-day basis, through which they can integrate Pixus. So one is simplifying AI, second is the transparency part," he said.
According to Val, transparency is woven into every automated decision the system makes, including what actions were taken, why, and how they impact performance. "We just want the marketers to have a framework where AI is in a very simplified format – where for them it's just a co-pilot. They still do their day-to-day operations, but the velocity with which they can do digital marketing significantly increases," he said.
When it comes to the impact of generative AI more broadly, Val sees significant change unfolding in the relationship between data scientists and marketing teams. "Right now since there is enough awareness about AI, there is a concept of data science team coming into the picture for running digital marketing campaigns where they're customising models," he said. "We already have a framework where we have a workbench for a data scientist and a workbench for a marketer – both can be integrated on a single dashboard."
Val also emphasised advances in automatically generating creative content tailored to historical performance, audience, and even product libraries. "Every brand right now, all the larger brands… they are sitting on a goldmine of data on Google and Meta … they already have historical data in terms of what type of ad creatives have worked or not worked," he said. "What we are working on building is a system that can crawl through the entire historical data, and going forward you just type in, 'generate an ad for an audience 25 to 35, target it to be in the geography of the United States', and the rest – language, theme, colour, background, call to action, all of it is automatically picked. The product is fetched from the product library, it can be automatically created, and the ad also will be linked to performance data."
On the biggest challenges marketers currently face, Val pointed out economic headwinds and shifting priorities. "The funding market is dry, and everyone has shifted their focus from growth to profitability… That's where a platform like us comes into the picture," he said. "Our pitch is very simple: you test out Pixus head-to-head against any manual system that you're running; we guarantee that your performance will improve by 10 to 15% in the first 8 to 10 weeks of deploying the platform."
But it's not just about improving returns. Val said Pixus helps brands optimise their internal resources as well. "Can I optimise my internal bandwidth – the same marketing team that I have internally, can we optimise their time so that they can focus on more strategic work rather than the executional day-to-day tasks?" he said.
He also claimed Pixus gives brands a unique advantage during peak periods such as Black Friday and Cyber Monday, when campaigns traditionally scale up by several multiples. "Those scale-ups up until now were happening manually – let's shift more dollars towards Facebook or Google. We have an intelligent system where through historical data we know which days, which seasonality, based on time of day, day of week – do weekends perform better than weekdays? There could be days when Facebook is doing better than Google, so shift budget from Google to Facebook in real-time."
Ultimately, Val argued, Pixus creates intellectual property for each brand it works with by training their AI on the brand's specific marketing data, resulting in "an AI that is being trained, which then becomes an IP for every brand."
For businesses interested in engaging with Pixus, Val said there are two main routes. "On the website, we have a 'Book a Demo' CTA or a button that they can approach directly, or if it's a large customer, they can directly email me and I can put them in touch with our sales team," he said.
Val finished by reflecting on the company's ambitions. "We just want to augment operations, simplify AI, and help marketers focus on strategy, not just execution," he said. "Absolutely, thank you so much for having me on."